This invention relates to a practice for determining the spatial location of the points on an object surface and, in particular, to a practice utilizing photographic two-dimensional records for generating three-dimensional information for defining object surface points.
Various techniques have been developed for three-dimensional reproduction of object surfaces. In one technique, referred to as photogrammetry, records of the object surface are simultaneously made by two spatially separated cameras of known location. These records are then examined to identify the corresponding points in each record representing a particular object feature. With this information and the known information regarding the locations and magnifications of the cameras, the spatial location of the particular object feature can be determined by simple triangulation. This prodedure is then repeated for other recorded object features until substantial reproduction of the object surface is at hand.
Photogrammetry is advantageous in situations where object motion is encountered, since the records of the bject can be made instantaneously and, therefore, are not affected by object motion. Typically, such instantaneous recording is carried out by utilizing a high speed strobe light to illuminate the object or, conversely, by using continuous illumination with cameras enabled by high speed synchronized shutters.
The aforesaid benefits of the photogrammetric procedure are, however, countered by the extremely long time that it takes to complete the operation. This, in turn, is due to the need to determine corresponding record points representative of the same object feature. Furthermore, in many situations, the degree of unique surface detail is limited, as in the case of a smooth surface, making it extremely difficult, if not impossible, to recognize features either by operator identification or by computer correlation of record points corresponding to object features.
Another technique developed for three dimensional reproduction eliminates the need for a lengthy recognition procedure. In this practice, as is disclosed in U.S. Pat. No. 3,866,052, assigned to the same assignee hereof, spatially encoded radiation patterns are projected at the object in time succession and records are made at the time of each such projection. These records may be used to determine for each recorded surface point a unique area corresponding to a particular area of illumination (i.e., a unique group of records) which corresponds in turn to a particular small surface area or feature of the object. Correlation of record points with corresponding object features is thus through coding and the need for operator identification and/or computer correlation via feature shape is thereby eliminated. In this case, the recorded object points and the known projector and camera orientations and locations can be used with the geometric triangulation procedure to arrive at the spatial location of the object surface points. For example, when the spatially encoded radiant energy patterns approach a planar form, each surface point can be found by the intersection of a particular camera ray corresponding to a particular recorded point with a particular projected planar pattern. Similarly, when each of the encoded patterns approximates an illuminating ray, each particular surface point is found by the intersection of a particular camera ray and a particular illuminating ray.
While the projection of a succession of encoded radiant energy patterns eliminates the need for a time consuming record investigation, the making of the records in this case is not instantaneous, due to film transport time, scan time, etc. As a result, any motion of the object during recording scrambles or confuses the recorded object surface information, making it difficult to reproduce therefrom the object surface.
It is therefore an object of the present invention to provide a method of three dimensional surface reproduction which is both rapid and substantially insensitive to object motion.